ursa-cli
ursa is a SQL engine for graph analytics, used from the command line. It reads Parquet, CSV, JSON/NDJSON (.json, .jsonl, .ndjson) and Arrow files where they are, on your disk or in object storage, and runs graph algorithms over them with SQL. There is no server and nothing to load first. An edge table is a graph.
On Linux or macOS, install the prebuilt binary with the install script:
|
On Windows, in PowerShell:
irm https://cldixon.github.io/ursa/install.ps1 | iex
Or build it with Cargo:
The binary is called ursa. The prebuilt binaries for Linux, macOS and Windows are also on the releases page, with a SHA-256 checksum for each.
* FROM 'karate_edges.csv';
CREATE PROPERTY GRAPH karate
EDGE TABLES (ties SOURCE KEY (src) DESTINATION KEY (dst));
SELECT * FROM pagerank(karate, direction => 'both')
ORDER BY pagerank DESC
LIMIT 5;
Graphs are defined with SQL/PGQ CREATE PROPERTY GRAPH. Algorithms and traversals are table functions with named parameters. SELECT * FROM ursa_functions() lists them all. A .ursa database file keeps your views, graphs and the graph indexes Ursa builds, so the next session starts where you left off.
For scripts and programs, ursa -c "<sql>", ursa -f script.sql and SQL on stdin run and exit; a script may use the shell's dot-commands (.mode, .timer on, .read), as in DuckDB. --format json writes results for a program to read. ursa --skill prints a guide to writing Ursa SQL for coding agents, and ursa --help ends with where to start.
The engine is built on Apache DataFusion. It runs the same algorithms as Ursa's Python library (ursa-graph) and gives the same values.
- Documentation: https://cldixon.github.io/ursa/sql/
- Source: https://github.com/cldixon/ursa
Licensed under either of MIT or Apache-2.0, at your option.